SCIENCE CHINA Life Sciences

© The Author(s) 2014. This article is published with open access at link.springer.com life.scichina.com link.springer.com

*Corresponding author (email: chenjiong@nju.edu.cn)

SPECIAL TOPIC: Model animals and their applications April 2015 Vol.58 No.4: 379–389

• RESEARCH PAPER • doi: 10.1007/s11427-014-4786-z

In vivo RNAi screen identifies candidate signaling genes required for collective cell migration in Drosophila ovary

LUO Jun, ZUO JunTao, WU Jing, WAN Ping, KANG Di, XIANG Cong, ZHU Hong & CHEN Jiong*

Model Animal Research Center, and MOE Key Laboratory of Model Animals for Disease Study, Nanjing University, Nanjing 210061, China

Received July 9, 2014; accepted November 11, 2014; published online December 19, 2014

Collective migration of loosely or closely associated cell groups is prevalent in animal development, physiological events, and cancer metastasis. However, our understanding of the mechanisms of collective cell migration is incomplete. Drosophila bor-der cells provide a powerful in vivo genetic model to study collective migration and identify essential genes for this process. Using border cell-specific RNAi-silencing in Drosophila, we knocked down 360 conserved signaling transduction genes in adult flies to identify essential pathways and genes for border cell migration. We uncovered a plethora of signaling genes, a large proportion of which had not been reported for border cells, including Rack1 (Receptor of activated C kinase) and brk (brinker), mad (mother against dpp), and sax (saxophone), which encode three components of TGF-β signaling. The RNAi knock down phenotype was validated by clonal analysis of Rack1 mutants. Our data suggest that inhibition of Src activity by Rack1 may be important for border cell migration and cluster cohesion maintenance. Lastly, results from our screen not only would shed light on signaling pathways involved in collective migration during embryogenesis and organogenesis in general, but also could help our understanding for the functions of conserved human genes involved in cancer metastasis.

Drosophila, border cell migration, signaling pathway, TGF-β, Brk, Rack1, Src42A, Src64B

Citation: Luo J, Zuo JT, Wu J, Wan P, Kang D, Xiang C, Zhu H, Chen J. In vivo RNAi screen identifies candidate signaling genes required for collective cell migration in Drosophila ovary. Sci China Life Sci, 2015, 58: 379–389, doi: 10.1007/s11427-014-4786-z

Cell migration is critical for embryonic development, adult wound healing, and immune system function. Understand-ing the mechanism of cell migration under physiological and developmental conditions can help better understand the underlying cause of abnormal cell migration under pathological conditions such as tumor metastasis and in-flammation. While some cells move singly, others migrate collectively as groups [1]. During gastrulation in the embryo and in epithelial sheet migration during wound healing, cells often migrate collectively. Collective migration has also been observed during tumor metastasis [13]. During col-lective migration, how individual cells receive and integrate outside signals to coordinate group migration is poorly un-

derstood. Studies in the Drosophila border cell migration have provided insight into the mechanisms of collective migration [2,4]. Border cell cluster is a specialized group of cells that migrate during oogenesis [4,5]. The Drosophila ovary contains strings of developing egg chambers. Each egg chamber is composed of an oocyte and 15 nurse cells surrounded by a monolayer of follicle cell epithelium [4,6]. Border cells are first specified and selected out from the anterior follicle cell epithelium at early stage 9 of oogenesis. The polar cells, which are a specialized pair of follicle cells at the anterior end of the egg chamber, recruit 4–8 cells surrounding the polar cells to become border cell cluster. Border cells detach from the follicle cell epithelium and migrate between the nurse cells over ~150 µm and reach the oocyte border by stage 10 (Figure 1A).

380 Luo J, et al. Sci China Life Sci April (2015) Vol.58 No.4

Figure 1 Overview of the screen. A, A diagram of border cell migration, the method of quantification of border cell migration delay and the classification of delay phenotype. The extent of migration for all stage 10 egg chambers examined was categorized as 0% (no migration), 25%, 50%, 75%, or 100% (reaching the border) for all quantitative analysis of border cell migration. B, Pie chart results from the RNAi screen. C, Typical samples from 0%, 25%, 50%, 75% and 100% border cell migration categories. D, Quantification of border cell migration for RNAi of selected genes. These genes have been report-ed to be required for border cell migration [712]. E, Quantification of border cell migration for RNAi of previously unreported signaling genes.

So far, at least five known signaling pathways have been extensively studied and shown to be required for different aspects of border cell migration. Steroid hormone signaling pathway coordinates the timing of migration [12,13]. Ecdy-sone activates the heterodimer of the nuclear hormone re-ceptor ecdysone receptor (EcR) and ultraspiracle (USP), both of which are necessary for border cell migration [12]. JAK/STAT signaling pathway specifies border cell identity and promotes expression of migratory genes [1417]. The cytokine in JAK/STAT signaling, unpaired (Upd), is se-creted by the two polar cells. Only 48 cells closest to the polar cells differentiate to border cells, thus the number of border cells is controlled by STAT levels. PDGF and VEGF-related receptor (PVR) and epidermal growth factor receptor (EGFR) signalings guide the chemotactic move-ment of border cells to the oocyte, in response to gradients of extracellular guidance factors including PVF and EGFR ligands [810,18]. The guidance function of PVR is redun-dant with that of EGFR. Jun amino-terminal kinase (JNK) signaling pathway regulates cell adhesion between border cells to maintain the cluster during migration [19,20]. Down-regulation of JNK signaling causes the cohesive cluster of border cells to dissociate [19]. The GTP exchange factor (GEF) Myoblast city and the small GTPase Rac are required for border cell migration and they act downstream of PVR [9]. Recently, Hedgehog signaling has been report-

ed to interact with the small GTPase Rac and the polarity protein Par1 in border cell migration, but the regulating mechanism is unknown [21]. Though these five signaling pathways have been reported to regulate border cell migra-tion, many gaps remain in our understanding of signaling regulation. Whether other signaling pathways contribute to border cell migration and how all of these signals are inte-grated to regulate border cell migration are largely un-known.

The TGF-β signaling plays major roles in many devel-opmental processes in Drosophila, but its roles in border cell migration are not reported. The major ligand of the TGF-β signaling in Drosophila is Decapentaplegic (Dpp) [22]. In the Drosophila ovary, Dpp is expressed in the ante-rior 20–30 follicle cells from stage 8 egg chambers. In late stage egg chambers, Dpp is expressed in all anterior follicle cells, including the stretched cells, the border cells, and the centripetal cells. Dpp protein forms a gradient from anterior to posterior in the follicle cells surrounding the oocyte [23]. Brinker is the target protein of the Drosophila Dpp mor-phogen signaling pathway. In Drosophila, two mechanisms have been proposed in the activation of Dpp target genes. One is through Dpp-dependent receptor activation which induces phosphorylation of Mad. Then Mad forms a heter-omeric complex with Medea (Med) in the nucleus and func-tions as an activator [24]. The other one is through inhibit-

Luo J, et al. Sci China Life Sci April (2015) Vol.58 No.4 381

ing Brinker (Brk), a transcriptional repressor [2527]. In fly wing development, Dpp is expressed in a narrow stripe across the anterior/posterior (A/P) boundary and Brk’s ex-pression pattern complements with that of the Dpp gradient. Brk is repressed by Dpp through a Schnurri-Mad-Med re-pressing complex [28]. On the other hand, Brk is also able to repress TGF-β target gene transcription by binding to Mad’s binding sites and competing with Mad [29,30]. Last-ly, Brk can recruit transcriptional co-repressors C-terminal binding protein (CtBP) and Groucho (Gro) via CtBP- and Gro-interaction motifs (CiM and GiM) [31,32].

PKC signaling pathway is conserved from yeast to hu-man. It modulates integrins and chemokine response, and regulates cell adhesion, migration, differentiation and pro-liferation [33]. The PKC anchoring protein Rack1 (receptor for activated C kinase 1) has been described to play a role in diverse processes based on cell culture studies, including cell adhesion, migration [3438], apoptosis, cell survival [3942], cell growth [43], and protein translation [4447]. Diverse protein-binding partners involved in key signaling pathways, such as PKC, Src, Integrin, IGF-I receptor, PP2A and FAK, are reported to bind to Rack1 in vitro [35,4854]. The loss-of-function phenotypes of Rack1 in Drosophila have been reported and they included reduction of the size of ovary and fewer germ cells per egg chamber in the Rack1 mutants, suggesting that Rack1 may function in cell divi-sion [55]. But the in vivo function of Rack1 in cell migra-tion is unknown.

Src proteins couple RTKs (receptor tyrosine kinases), in-tegrin and GPCR (G-protein coupled receptor) to intracel-lular signaling pathways and regulate cell division and cell motility [56]. Interaction between Rack1 and Src has been implicated in the regulation of cell proliferation and migra-tion in cell culture studies [36,43]. Rack1 has been reported to bind to Src [50,57,58], which in turn phosphorylate Rack1 itself, resulting in eventual negative regulation of Src activity [36,57,59]. Src activation promotes cell cycle entry, cytoskeletal rearrangements, and alterations in cell adhesion [60]. Activated Src induces the EMT (epithelium to mesen-chyme transition) process through tyrosine phosphorylation of adherens junction components (the cadherin/catenin sys-tem) or acting by phosphorylating protein regulators of ad-herens junction [61]. The Drosophila genome contains two Src family kinases, Src42A and Src64B, which are func-tionally similar to their mammalian counterparts [62,63]. Src42A is the closest relative of vertebrate Src in Drosoph-ila and functions redundantly to Src64B, and it is found to genetically interact with DE-cadherin and Armadillo [64]. In Drosophila oogenesis, strong Src42A expression is evi-dent in invading and migrating border cells [64]. However, the function of Src in border cell migration is unknown.

Loss-of-function screens, including EMS induced muta-genesis, had been performed in the past, and they identified many of the important genes required for border cell migra-tion, such as Slbo, Pvf1, Apontic, Par1, Taiman, Jing, Psidin,

Stat92E, Puckered, and Sec3 [7,12,14,15,6569]. To bypass early lethality (earlier requirement of the genes essential for the survival of embryos and larva), these screens were often clonally (mosaic) based and utilized FRT/FLP technique. However, one of the drawbacks of this approach is that mo-saic clones encompassing entire border cell clusters are not always easily obtained. Furthermore, it is time-consuming to map EMS-induced mutation and to clone the affected genes. Alternatively, tissue-specific RNAi is a fast and effi-cient way to perform a loss-of-function screen in Drosophi-la. Here, we performed an in vivo RNAi screen for candi-date genes in major signaling pathways required for border cell migration. Overall, we isolated 111 candidate genes required for border cell migration from this screen. Among them, Rack1 and three TGF-β signaling components, sax, mad and brk were characterized in more details.

1 Materials and methods

1.1 Drosophila genetics

Flies were cultured following standard procedures at 25°C except for RNAi experiments at 29°C. Most of the RNAi stocks were obtained from Japan NIG-fly stock center and some were obtained from Vienna Drosophila Resource Center (VDRC). Other strains were obtained from the Bloomington Drosophila Stock Center. To generate UAS-Rack1.GFP transgenic line, we subcloned a full-length cDNA of the Rack1 gene into a modified pUAST-attB vec-tor with C-terminal GFP tagging. The resulting UAS- Rack1.GFP construct was injected into embryos according to standard procedures. Mutant FRT clones were induced using hs-FLP. Flies were heat shocked for 1 h per day at 37°C for 3 d before eclosion, then fed with yeast and dis-sected 2–3 d after the last heat shock. slbo-Gal4 specifically drives expression in border cells, and has been extensively used by many labs to over-express genes in the border cells [70]. UAS-Dicer2 was used along with UAS-Candidate gene RNAi to enhance RNAi efficiency. Culture temperature of 29°C was used to enhance the activity of the UAS/Gal4 system without affecting survival. Lastly, UAS-GFP was used to mark the border cell clusters.

1.2 Screening system

The scheme of the screening was described below. Crossing of virgin females (UAS-Dcr2;slbo-Gal4,UAS-GFP/CyO) with 2–3 males from UAS-RNAi lines were set at 25°C. UAS-Dcr2;slbo-Gal4,UAS-GFP/CyO was outcrossed to w1118 and used as controls. Two to three days post eclo-sion, female adult flies from F1 generation were shifted to 29°C to enhance Gal4 activity and RNAi expression. After 3 d, ovaries of the female flies were dissected in PBS, fixed in 7% paraformaldehyde for 10 min at room temperature, rinsed in PBS+0.3% Triton X-100 and PBS. Fixed ovaries

382 Luo J, et al. Sci China Life Sci April (2015) Vol.58 No.4

were manually dissociated in 80% glycerol. UAS-GFP flu-orescence was used to visualize border cells in dissociated ovaries. Analysis of border cell migration was performed with an Olympus BX51 fluorescent microscope.

The quantification and categories of the phenotype are described below. The extent of migration for all stage 10 egg chambers examined was categorized as 0% (no migra-tion), 25%, 50%, 75%, or 100% (reaching the border) for quantitative analysis of border cell migration (Figure 1A and C). The D value is calculated as the number of border cell migration delay egg chambers divided by the total number of examined egg chambers. The D value between 0.4 and 0.6 was considered weak migration delay phenotype; the D value between 0.6 and 0.8 was the intermediate phe-notype; the D value between 0.8 and 1 was considered the severe phenotype (Figure 1A). NI was calculated as the number of stage 10 egg chambers exhibiting border cell non-invasion divided by the total number of stage 10 egg chambers exhibiting migration delay. If the NI0.4, we classified the genes into the invasion defective group.

1.3 Immunohistochemistry and microscopy

Ovary dissection was carried out in phosphate-buffered sa-line (PBS) and then fixed in Devitellinizing buffer (7% formaldehyde) and heptane (Sigma) mixture (1:6) for 10 min. After washes in PBS, ovaries were incubated in blocking solution (PBT, 10% goat serum) for 30 min and then stained overnight at 4°C. Primary antibodies and their concentra-tions were as follows: rat anti-DEcad (DCAD2, 1:50, DSHB). After washes in PBT, ovaries were incubated with secondary antibodies (Jackson ImmunoResearch) for 2 h at room temperature. F-actin was labeled by rhodamine phal-loidin (1:100; Sigma). DNA was labeled by DAPI (1:1000; Sigma). Confocal images were obtained using a Leica TCS SP5 II or an Olympus FV1000 confocal microscope.

2 Results

2.1 Overview of the RNAi screen

To identify novel candidate signaling molecules for border cell migration, we conducted a border cell-specific RNAi silencing screen of signaling pathway genes that we identi-fied from the GO (Gene Ontology) Term list in the FlyBase website (http://flybase.org) (Table S1 in Supporting Infor-mation).

The components of major signaling pathways in Dro-sophila were selected for this RNAi screen. We restricted our screen to 572 UAS-RNAi lines corresponding to 360 genes in a variety of major signal transduction pathways, including Wnt, TGF-β, Hh, Integrin, Toll, G protein cou-pled receptor (GPCR), Notch, small GTPases, stress, and apoptosis signaling pathways. About half of these genes had two corresponding RNAi lines in this screen. We used D

(delay) value to quantify the border cell migration delay phenotype (Figure 1A). D value reflected the proportion of the migration delay border cell clusters. A D value of 1.0 means that 100% of stage 10 egg chambers examined dis-play various extents of migration delay, which is character-ized by failure of border cells reaching the border of oocyte. Of the 360 genes, we found about 70% of genes had no sig-nificant migration delay (D value below 0.4) when knocked down, 15% had weak migration delay phenotype (0.4<D <0.6), 8% had intermediate phenotype (0.6<D<0.8) and 7% had severe phenotype (0.8<D<1.0; Figure 1B). Next, we subdivided the migration delay genes into two groups by the border cell cluster’s detachment or invasion ability. We used NI (noninvasion) value to measure the border cell cluster’s detachment or invasion defect (Figure 1A). NI value reflected the proportion of the no migration clusters among migration delay border cell clusters. An NI value of 1.0 means that all the migration delayed stage 10 egg chambers display a complete lack of migration by border cells. These border cells still remained attached to the ante-rior end of egg chamber and appeared not to invade and initiate migration between underlying nurse cells. Therefore, we would define such a defect as a non-invasion phenotype. An NI value equal or greater than 0.4 means a significant invasion defect. From comparing the D value and the NI value, we found that significant invasion defects (NI0.4) were often associated with severe migration delay pheno-types (0.8<D<1.0), and most of the transcription regulator encoding genes isolated from this screen showed invasion defects (NI0.4) when knocked down (Table 2), suggesting that the transition from border cells’ non-invasive state to the invasive (early migratory) state requires much transcrip-tional activity.

Overall, we uncovered 111 candidate genes required for border cell migration from this screen (Table 1). Six of these genes were previously found to affect border cell mi-gration, including slbo, pvr, egfr, Rac1, Rac2 and Ecdysone receptor, validating the efficacy of our screen. (Figure 1D). For example, 87.5% of egfr RNAi stage 10 egg chambers displayed border cell migration delay (Figure 1D). Of those with migration delay, three out of four stage 10 egg cham-bers are in 75% migration category, suggesting that the EGFR signaling mainly promotes the later stage of border cell migration, which is consistent with the previous report [8]. More importantly and interestingly, we have found for the first time that RNAi of some key components of various signaling pathways caused migration delay, such as dsh (disheveled), Rh2 (Rhodopsin 2), dorsal, p38b and pelle (pll) (Figure 1E). Dishevelled (Dsh) is a key component of Wnt signaling. Dorsal and Pelle are components of Toll signal-ing. Dorsal is a transcriptional factor and Pelle is a Ser-ine/Threonine protein kinase. p38b is a MAPK (mito-gen-activated protein kinase). Rh2 is a G-protein coupled receptor. Furthermore, diverse receptors, transcription reg-ulators and kinases are implicated for the first time to func-

Luo J, et al. Sci China Life Sci April (2015) Vol.58 No.4 383

Table 1 Summary of the selected genes and the screen resulta)

Signaling Wnt Hh TGFb Toll Notch GPC1 Stress Apo2 NT3 SG4 Other Total

Gene tested 28 23 23 28 17 12 16 15 20 24 154 360

Gene (D40%) 11 7 6 11 8 8 4 4 2 6 44 111

Gene (NI40%) 7 4 3 6 6 2 3 3 2 1 23 60

a) 1, GPC: G protein coupled signaling; 2, Apo: apoptosis signaling; 3, NT: neurotransmitter transporter; 4, SG: small GTPase.

Table 2 Molecular function categories of the uncovered candidate genes

Class Receptor Transcription

regulator Kinase

Small GTPase

Ligand mRNA binding

Protein binding

Transporter

Other Total

Gene (D40%) 18 20 11 6 6 3 15 3 29 111

Gene (NI40%) 6 16 8 0 2 2 6 3 17 60

tion in border cell migration (Table 2). Below we described two signaling pathways recovered from the RNAi screen, the TGF-β signaling and Rack1-mediated PKC signaling.

2.2 RNAi knockdown of brk, mad or sax affects border cell migration

The TGF-β signaling components, Brk (Brinker), Mad (Mothers against dpp), and Sax (Saxophone), were uncov-ered from the RNAi screen. Brk is a transcriptional re-pressor and a target protein of the Drosophila Dpp mor-phogen signaling pathway. Inhibiting Brk can activate Dpp target genes. We found that knockdown of brk caused very strong border cell migration defect (Figure 2C and F). In brk RNAi, all of the border cell clusters (D=1.0) failed to reach the border and most of them (NI=0.82) had invasion defect, suggesting that brk was required for border cell in-vasion process probably by inhibiting the expression of Dpp target genes. Furthermore, the high values of both D and NI suggest that brk may be required for the initial invasive stage of border cell migration. Mad is a receptor-regulated Smad (R-Smad) in Drosophila. In mad RNAi, about 60% (D=0.56) stage 10 egg chambers showed border cell migra-tion defect (Figure 2D and F), and the NI value (0.19) did not indicate a significant invasive defect, suggesting that mad is required in the mid-migratory phase of the border cell migration rather than in the initial phase. We did not find that other Smads were required for border cell migra-tion in this screen, implying that border cell migration was probably regulated by non-canonical signaling functions of Mad. Sax is a type I receptor in Drosophila, which can propagate the signal through phosphorylation of the Smad proteins. In sax RNAi, 63% (D=0.63) of stage 10 egg chambers showed border cell migration defect (Figure 2E and F), while NI value is low (0.27), suggesting that like mad, sax is also required in the mid-migratory phase of the

migration process. Taken together, results from brk, mad and sax indicate that keeping a proper level of TGF-β sig-naling during different stages (invasive and migratory stag-es) is essential for border cell migration. Over-activation of TGF-β signaling by inhibiting Brk during the initial inva-sive stage and down-regulation of TGF-β signaling by Mad deficiency caused distinct migration defects.

2.3 Rack1 is required for border cell cluster migration and organization

Another candidate gene that we further analyzed in details was Rack1. Interaction between Rack1 and Src has been implicated in the regulation of cell proliferation and migra-tion in cell culture studies [36,43]. Rack1 has been previ-ously reported to negatively regulate Src activity [36,57,59]. In our RNAi screen, we found knockdown of Rack1 caused moderate delay of border cell migration (D=0.29, NI=0; Figure 3C and H), suggesting that Rack1 was required for the migratory process of border cell cluster, rather than the early invasion and delamination stage of border cells. To validate the result from RNAi knockdown, we generated Rack1 mutant clones in border cells. Three Rack1 mutant alleles have been reported [55]. Rack1EY128 is a null allele in which a P-element is inserted 53 bp upstream of the transla-tion start site. Rack1EE is a hypomorphic allele which ex-presses S81F mutant protein at a much lower level. Rack11.8 is a null allele which changes glutamine 6 to a stop codon. About 30% of the mosaic border cell clusters containing Rack11.8, Rack1EE or Rack1EY128 homozygous clones dis-played migration delay (Figure 4A–C′″ and E). Thus, results from both RNAi and genetic mosaic analysis indicate that Rack1 is required for border cell migration.

In Rack1 knockdown or Rack1 mutant border cells, we found abnormal arrangement of border cell cluster with one or more individual cells trailing behind the main cluster

384 Luo J, et al. Sci China Life Sci April (2015) Vol.58 No.4

Figure 2 TGF-beta signaling components brk, mad and sax are candidate genes required for border cell migration. AA′″, Wild-type mid stage 9 egg chamber showing border cell migrating through the nurse cells. BB′″, Wild-type stage 10 egg chamber showing border cell reaching the border between nurse cells and oocyte. CC′″, Border cell migration is severely delayed when brk is knocked down. DD′″, Border cell migration is delayed when mad is knocked down. EE′″, Border cell migration is delayed when sax is knocked down. F, Quantitation of border cell migration. The y-axis denotes the percent-age of stage 10 egg chambers examined for each genotype that exhibits each degree of migration, as represented by the five color-coded bars for each geno-type. Arrows indicate the border cells. F-actin is labeled by rhodamine phalloidin. Scale bar, 50 µm.

(Figure 3C–E′″). The expression pattern of the adherens junction protein E-cadherin was affected in Rack1 knock-down border cell clusters (Figure 5). The dissociated cluster phenotype and the affected adherens junctions suggest that Rack1 functions in regulating cell adhesion between border cells, which has not been previously reported. Activated Src has been reported to induce EMT process through tyrosine phosphorylation of adherens junction components (the E-cadherin/β-catenin system) or by phosphorylating regula-tors of adherens junctions [61]. We found that overexpres-sion of Src42A.CA (constitutively active form) or Src64B

caused strong migration delay and abnormal arrangement (dissociated cluster) of border cells, which was similar to the Rack1 loss-of-function phenotype (Figure 3F–H). These results suggested that Src activity was regulated by Rack1 in border cells. Our data suggest that in Rack1 deficient border cells, Src activity was not inhibited by Rack1 and the activated Src induced reduction of cell-cell junctions be-tween border cells, which resulted in dissociated border cells. Therefore, we hypothesize that Rack1 negatively reg-ulates Src activity to stabilize cell-cell junctions and pro-motes cell migration. Lastly, we generated the UAS-Rack1-

Luo J, et al. Sci China Life Sci April (2015) Vol.58 No.4 385

Figure 3 Rack1 is required for border cell migration and cluster cohesion. AA′″, Wild-type stage 9 egg chamber shows border cell migration. BB′″, Wild-type stage 10 egg chamber shows border cells reaching the border. CE′″, Border cell cluster cohesion is disrupted when Rack1 is knocked down. F, Border cell cluster arrangement is disrupted when Src42A.CA is overexpressed. G, Border cell migration is delayed when Src64B is overexpressed. H, Quan-titation of border cell migration for each genotype. Arrows indicate the border cells. F-actin is labeled by rhodamine phalloidin. Scale bar, 50 µm.

GFP transgenic fly to analyze the subcellular localization of Rack1. slbo-Gal4 was used to drive its expression in border cells. We found that Rack1-GFP was enriched in cell-cell junctions, suggesting that Rack1 was probably involved in cell-cell adhesion or cell-cell communication.

3 Discussion

Here, we present an in vivo RNAi border cell screen for a collection of Drosophila signaling genes. We uncovered a plethora of signaling genes, a large proportion of which had

not been previously reported for border cell migration. Among these, 18 genes encode for receptors, and they in-clude Wnt signaling receptor fz3, type I TGFβ receptor sax, type II TGFβ receptor wit and G protein coupled receptors Rh2 and GRHR, suggesting that border cell migration could be regulated by diverse signaling from extracellular factors. According to our data, there is not any single pathway that has all of their components isolated from the screen. Firstly, this implies that slbo-Gal4 driven RNAi may not be sensi-tive enough to isolate all genes in our case. After all, the effective time window for slbo-Ga4 expression only last for about 6 h, spanning from early stage 9 to late stage 9. Thus,

386 Luo J, et al. Sci China Life Sci April (2015) Vol.58 No.4

Figure 4 Validation of the Rack1 knockdown phenotype. AC′″, Border cell migration is delayed for border cell cluster including Rack1EY128 mutant clones (AA′″), Rack1EE mutant clones (BB′″), or Rack11.8 mutant clones (CC′″). Arrows indicate the mutant border cells. D, The expression pattern of UAS-Rack1.GFP in border cells. E, Quantitation of border cell migration in Rack1 mutant clones. F-actin is labeled by rhodamine phalloidin. Mutant clones are marked by the lack of GFP. Scale bar, 50 µm in AC′″, and 10 µm in D.

during this small time window, the level of knockdown may not be effective enough for all genes tested. Secondly, this implies that border cell migration could be regulated by novel function of individual signaling components or by the non-canonical signaling functions of these components.

Transcription regulators have been reported to be critical for border cell migration, such as Slbo, STAT, Taiman, Apontic and Abrupt [7,12,13,15,68]. From our screen, we found that knockdown of 20 transcription regulator encod-ing genes caused border cell migration delay and 16 of them also caused border cell cluster invasion defects, such as brk, CtBP, CHES-1-like, and osa. CHES-1-like is a checkpoint suppressor homologue and is related to Foxn2 and Foxn3, the mouse counterpart of human CHES1 [71]. Osa is a chromatin-remodeling protein and interacts with Cyclin E in Drosophila eye imaginal discs [72]. Most of them have not been reported to be required for border cell migration. This result suggests that transcription regulators mainly drive the detachment or invasive stage of border cell clusters.

One pathway we recovered was the Sax, Mad and Brk

from TGF-β signaling. TGF-β is an inducer of EMT process [73]. In cancer progression, EMT was associated with tumor invasiveness. During border cell migration, the invasion of border cell cluster is an EMT-like process. From the RNAi screen, we found that TGF-β signaling may also contribute to the invasion of border cell clusters. However, not all TGF-β signaling pathway components are uncovered in the RNAi screen, suggesting that TGF-β signaling regulates border cell migration probably through a non-canonical fashion.

Another pathway we identified was the Rack1-Src path-way. Border cell migration analysis in Drosophila shows that RNAi knockdown of Rack1 or overexpression of Src results in border cell migration and cluster cohesion defects. Consistently, Rack1 loss-of-function mosaic clones in bor-der cell clusters resulted in similar phenotypes, suggesting that the inhibition of Src activity by Rack1 in border cells is critical for migration and cluster cohesion, probably through regulating cell-cell adhesion between border cells or be-tween border cells and nurse cells.

Luo J, et al. Sci China Life Sci April (2015) Vol.58 No.4 387

Figure 5 E-cadherin staining is affected in Rack1 RNAi border cell clusters. AA″, E-cadherin expression pattern in the wild-type border cell cluster. BC″, E-cadherin level is significantly reduced in border cell cluster that is split up (BB″) or in cluster that is still coherent (CC″). DAPI labels the nuclei of border cells and the large nuclei of the surrounding nurse cells. Arrows indicate the split border cell in CC″ and the border cell that lag behind in BB″. Scale bar, 10 µm.

Lastly, we demonstrate that RNAi-based forward screen is a fast and efficient way to uncover candidate genes for border cell migration. However, all these candidate genes still need to be further validated and confirmed in the future by using loss-of-function mutations in mosaic analysis as we have done for Rack1.

We thank Bloomington Drosophila Stock Center, NIG-Fly, and Vienna Drosophila Resource Center for fly stocks. This work was supported by grants from the National Natural Science Foundation of China (31271488, 31171335, 31071219) to Chen Jiong.

1 Friedl P, Gilmour D. Collective cell migration in morphogenesis, re-generation and cancer. Nat Rev Mol Cell Biol, 2009, 10: 445–457

2 Yilmaz M, Christofori G. Mechanisms of motility in metastasizing cells. Mol Cancer Res, 2010, 8: 629–642

3 Friedl P, Locker J, Sahai E, Segall JE. Classifying collective cancer cell invasion. Nat Cell Biol, 2012, 14: 777–783

4 He L, Wang X, Montell DJ. Shining light on Drosophila oogenesis: live imaging of egg development. Curr Opin Genet Dev, 2011, 21: 612–619

5 Montell DJ, Yoon WH, Starz-Gaiano M. Group choreography: mechanisms orchestrating the collective movement of border cells. Nat Rev Mol Cell Biol, 2012, 13: 631–645

6 Spradling AC. Germline cysts: communes that work. Cell, 1993, 72: 649–651

7 Montell DJ, Rorth P, Spradling AC. Slow border cells, a locus re-quired for a developmentally regulated cell migration during oogene-sis, encodes Drosophila C/EBP. Cell, 1992, 71: 51–62

8 Duchek P, Rorth P. Guidance of cell migration by EGF receptor sig-naling during Drosophila oogenesis. Science, 2001, 291: 131–133

9 Duchek P, Somogyi K, Jekely G, Beccari S, Rorth P. Guidance of cell migration by the Drosophila PDGF/VEGF receptor. Cell, 2001, 107: 17–26

10 McDonald JA, Pinheiro EM, Montell DJ. PVF1, a PDGF/VEGF homolog, is sufficient to guide border cells and interacts genetically with Taiman. Development, 2003, 130: 3469–3478

11 Murphy AM, Montell DJ. Cell type-specific roles for Cdc42, Rac, and RhoL in Drosophila oogenesis. J Cell Biol, 1996, 133: 617–630

12 Bai J, Uehara Y, Montell DJ. Regulation of invasive cell behavior by taiman, a Drosophila protein related to AIB1, a steroid receptor coac-tivator amplified in breast cancer. Cell, 2000, 103: 1047–1058

13 Jang AC, Chang YC, Bai J, Montell D. Border-cell migration re-quires integration of spatial and temporal signals by the BTB protein Abrupt. Nat Cell Biol, 2009, 11: 569–579

14 Silver DL, Montell DJ. Paracrine signaling through the JAK/STAT

388 Luo J, et al. Sci China Life Sci April (2015) Vol.58 No.4

pathway activates invasive behavior of ovarian epithelial cells in Drosophila. Cell, 2001, 107: 831–841

15 Beccari S, Teixeira L, Rorth P. The JAK/STAT pathway is required for border cell migration during Drosophila oogenesis. Mech Dev, 2002, 111: 115–123

16 Ghiglione C, Devergne O, Georgenthum E, Carballes F, Medioni C, Cerezo D, Noselli S. The Drosophila cytokine receptor Domeless controls border cell migration and epithelial polarization during oo-genesis. Development, 2002, 129: 5437–5447

17 Silver DL, Geisbrecht ER, Montell DJ. Requirement for JAK/STAT signaling throughout border cell migration in Drosophila. Develop-ment, 2005, 132: 3483–3492

18 McDonald JA, Pinheiro EM, Kadlec L, Schupbach T, Montell DJ. Multiple EGFR ligands participate in guiding migrating border cells. Dev Biol, 2006, 296: 94–103

19 Llense F, Martin-Blanco E. JNK signaling controls border cell cluster integrity and collective cell migration. Curr Biol, 2008, 18: 538–544

20 Melani M, Simpson KJ, Brugge JS, Montell D. Regulation of cell adhesion and collective cell migration by hindsight and its human homolog RREB1. Curr Biol, 2008, 18: 532–537

21 Geisbrecht ER, Sawant K, Su Y, Liu ZC, Silver DL, Burtscher A, Wang X, Zhu AJ, McDonald JA. Genetic interaction screens identify a role for hedgehog signaling in Drosophila border cell migration. Dev Dyn, 2013, 242: 414–431

22 Padgett RW, St Johnston RD, Gelbart WM. A transcript from a Dro-sophila pattern gene predicts a protein homologous to the transform-ing growth factor-beta family. Nature, 1987, 325: 81–84

23 Twombly V, Blackman RK, Jin H, Graff JM, Padgett RW, Gelbart WM. The TGF-beta signaling pathway is essential for Drosophila oogenesis. Development, 1996, 122: 1555–1565

24 Wisotzkey RG, Mehra A, Sutherland DJ, Dobens LL, Liu X, Dohrmann C, Attisano L, Raftery LA. Medea is a Drosophila Smad4 homolog that is differentially required to potentiate DPP responses. Development, 1998, 125: 1433–1445

25 Campbell G, Tomlinson A. Transducing the Dpp morphogen gradient in the wing of Drosophila: regulation of Dpp targets by brinker. Cell, 1999, 96: 553–562

26 Jazwinska A, Kirov N, Wieschaus E, Roth S, Rushlow C. The Dro-sophila gene brinker reveals a novel mechanism of Dpp target gene regulation. Cell, 1999, 96: 563–573

27 Minami M, Kinoshita N, Kamoshida Y, Tanimoto H, Tabata T. brinker is a target of Dpp in Drosophila that negatively regulates Dpp-dependent genes. Nature, 1999, 398: 242–246

28 Muller B, Hartmann B, Pyrowolakis G, Affolter M, Basler K. Con-version of an extracellular Dpp/BMP morphogen gradient into an in-verse transcriptional gradient. Cell, 2003, 113: 221–233

29 Saller E, Bienz M. Direct competition between Brinker and Drosoph-ila Mad in Dpp target gene transcription. EMBO Rep, 2001, 2: 298–305

30 Kirkpatrick H, Johnson K, Laughon A. Repression of dpp targets by binding of brinker to mad sites. J Biol Chem, 2001, 276: 18216–18222

31 Upadhyai P, Campbell G. Brinker possesses multiple mechanisms for repression because its primary co-repressor, Groucho, may be una-vailable in some cell types. Development, 2013, 140: 4256–4265

32 Hasson P, Muller B, Basler K, Paroush Z. Brinker requires two core-pressors for maximal and versatile repression in Dpp signalling. EMBO J, 2001, 20: 5725–5736

33 Spitaler M, Cantrell DA. Protein kinase C and beyond. Nat Immunol, 2004, 5: 785–790

34 Buensuceso CS, Woodside D, Huff JL, Plopper GE, O’Toole TE. The WD protein Rack1 mediates protein kinase C and integ-rin-dependent cell migration. J Cell Sci, 2001, 114: 1691–1698

35 Besson A, Wilson TL, Yong VW. The anchoring protein RACK1 links protein kinase Cepsilon to integrin beta chains. Requirements for adhesion and motility. J Biol Chem, 2002, 277: 22073–22084

36 Cox EA, Bennin D, Doan AT, O’Toole T, Huttenlocher A. RACK1 regulates integrin-mediated adhesion, protrusion, and chemotactic cell migration via its Src-binding site. Mol Biol Cell, 2003, 14:

658–669 37 Kiely PA, Leahy M, O’Gorman D, O’Connor R. RACK1-mediated

integration of adhesion and insulin-like growth factor I (IGF-I) sig-naling and cell migration are defective in cells expressing an IGF-I receptor mutated at tyrosines 1250 and 1251. J Biol Chem, 2005, 280: 7624–7633

38 Kiely PA, O’Gorman D, Luong K, Ron D, O’Connor R. Insulin-like growth factor I controls a mutually exclusive association of RACK1 with protein phosphatase 2A and beta1 integrin to promote cell mi-gration. Mol Cell Biol, 2006, 26: 4041–4051

39 Sang N, Severino A, Russo P, Baldi A, Giordano A, Mileo AM, Pag-gi MG, De Luca A. RACK1 interacts with E1A and rescues E1A-induced yeast growth inhibition and mammalian cell apoptosis. J Biol Chem, 2001, 276: 27026–27033

40 Choi DS, Young H, McMahon T, Wang D, Messing RO. The mouse RACK1 gene is regulated by nuclear factor-kappa B and contributes to cell survival. Mol Pharmacol, 2003, 64: 1541–1548

41 Mourtada-Maarabouni M, Kirkham L, Farzaneh F, Williams GT. Functional expression cloning reveals a central role for the receptor for activated protein kinase C 1 (RACK1) in T cell apoptosis. J Leu-koc Biol, 2005, 78: 503–514

42 Mamidipudi V, Cartwright CA. A novel pro-apoptotic function of RACK1: suppression of Src activity in the intrinsic and Akt pathways. Oncogene, 2009, 28: 4421–4433

43 Mamidipudi V, Dhillon NK, Parman T, Miller LD, Lee KC, Cart-wright CA. RACK1 inhibits colonic cell growth by regulating Src ac-tivity at cell cycle checkpoints. Oncogene, 2007, 26: 2914–2924

44 Shor B, Calaycay J, Rushbrook J, McLeod M. Cpc2/RACK1 is a ri-bosome-associated protein that promotes efficient translation in Schizosaccharomyces pombe. J Biol Chem, 2003, 278: 49119–49128

45 Nilsson J, Sengupta J, Frank J, Nissen P. Regulation of eukaryotic translation by the RACK1 protein: a platform for signalling mole-cules on the ribosome. EMBO Rep, 2004, 5: 1137–1141

46 Gerbasi VR, Weaver CM, Hill S, Friedman DB, Link AJ. Yeast Asc1p and mammalian RACK1 are functionally orthologous core 40S ribosomal proteins that repress gene expression. Mol Cell Biol, 2004, 24: 8276–8287

47 Sengupta J, Nilsson J, Gursky R, Spahn CM, Nissen P, Frank J. Iden-tification of the versatile scaffold protein RACK1 on the eukaryotic ribosome by cryo-EM. Nat Struct Mol Biol, 2004, 11: 957–962

48 Stebbins EG, Mochly-Rosen D. Binding specificity for RACK1 re-sides in the V5 region of beta II protein kinase C. J Biol Chem, 2001, 276: 29644–29650

49 Steele MR, McCahill A, Thompson DS, MacKenzie C, Isaacs NW, Houslay MD, Bolger GB. Identification of a surface on the be-ta-propeller protein RACK1 that interacts with the cAMP-specific phosphodiesterase PDE4D5. Cell Signal, 2001, 13: 507–513

50 Chang BY, Chiang M, Cartwright CA. The interaction of Src and RACK1 is enhanced by activation of protein kinase C and tyrosine phosphorylation of RACK1. J Biol Chem, 2001, 276: 20346–20356

51 Kiely PA, Sant A, O’Connor R. RACK1 is an insulin-like growth factor 1 (IGF-1) receptor-interacting protein that can regulate IGF-1-mediated Akt activation and protection from cell death. J Biol Chem, 2002, 277: 22581–22589

52 Kiely PA, Baillie GS, Barrett R, Buckley DA, Adams DR, Houslay MD, O’Connor R. Phosphorylation of RACK1 on tyrosine 52 by c-Abl is required for insulin-like growth factor I-mediated regulation of focal adhesion kinase. J Biol Chem, 2009, 284: 20263–20274

53 Kiely PA, Baillie GS, Lynch MJ, Houslay MD, O’Connor R. Tyro-sine 302 in RACK1 is essential for insulin-like growth fac-tor-I-mediated competitive binding of PP2A and beta1 integrin and for tumor cell proliferation and migration. J Biol Chem, 2008, 283: 22952–22961

54 Liliental J, Chang DD. Rack1, a receptor for activated protein kinase C, interacts with integrin beta subunit. J Biol Chem, 1998, 273: 2379–2383

55 Kadrmas JL, Smith MA, Pronovost SM, Beckerle MC. Characteriza-tion of RACK1 function in Drosophila development. Dev Dyn, 2007, 236: 2207–2215

Luo J, et al. Sci China Life Sci April (2015) Vol.58 No.4 389

56 Courtneidge SA. Role of Src in signal transduction pathways. The Jubilee Lecture. Biochem Soc Trans, 2002, 30: 11–17

57 Chang BY, Conroy KB, Machleder EM, Cartwright CA. RACK1, a receptor for activated C kinase and a homolog of the beta subunit of G proteins, inhibits activity of src tyrosine kinases and growth of NIH 3T3 cells. Mol Cell Biol, 1998, 18: 3245–3256

58 Chang BY, Harte RA, Cartwright CA. RACK1: a novel substrate for the Src protein-tyrosine kinase. Oncogene, 2002, 21: 7619–7629

59 Mamidipudi V, Chang BY, Harte RA, Lee KC, Cartwright CA. RACK1 inhibits the serum- and anchorage-independent growth of v-Src transformed cells. FEBS Lett, 2004, 567: 321–326

60 Abram CL, Courtneidge SA. Src family tyrosine kinases and growth factor signaling. Exp Cell Res, 2000, 254: 1–13

61 Frame MC. Newest findings on the oldest oncogene; how activated src does it. J Cell Sci, 2004, 117: 989–998

62 Dodson GS, Guarnieri DJ, Simon MA. Src64 is required for ovarian ring canal morphogenesis during Drosophila oogenesis. Develop-ment, 1998, 125: 2883–2892

63 Takahashi F, Endo S, Kojima T, Saigo K. Regulation of cell-cell contacts in developing Drosophila eyes by Dsrc41, a new, close rela-tive of vertebrate c-src. Genes Dev, 1996, 10: 1645–1656

64 Takahashi M, Takahashi F, Ui-Tei K, Kojima T, Saigo K. Require-ments of genetic interactions between Src42A, armadillo and shotgun, a gene encoding E-cadherin, for normal development in Drosophila. Development, 2005, 132: 2547–2559

65 Liu Y, Montell DJ. Identification of mutations that cause cell migra-

tion defects in mosaic clones. Development, 1999, 126: 1869–1878 66 Mathieu J, Sung HH, Pugieux C, Soetaert J, Rorth P. A sensitized

PiggyBac-based screen for regulators of border cell migration in Drosophila. Genetics, 2007, 176: 1579–1590

67 Wan P, Wang D, Luo J, Chu D, Wang H, Zhang L, Chen J. Guidance receptor promotes the asymmetric distribution of exocyst and recy-cling endosome during collective cell migration. Development, 2013, 140: 4797–4806

68 Starz-Gaiano M, Melani M, Wang X, Meinhardt H, Montell DJ. Feedback inhibition of Jak/STAT signaling by apontic is required to limit an invasive cell population. Dev Cell, 2008, 14: 726–738

69 McDonald JA, Khodyakova A, Aranjuez G, Dudley C, Montell DJ. PAR-1 kinase regulates epithelial detachment and directional protru-sion of migrating border cells. Curr Biol, 2008, 18: 1659–1667

70 Geisbrecht ER, Montell DJ. Myosin VI is required for E-cadherin-mediated border cell migration. Nat Cell Biol, 2002, 4: 616–620

71 Lee HH, Frasch M. Survey of forkhead domain encoding genes in the Drosophila genome: classification and embryonic expression patterns. Dev Dyn, 2004, 229: 357–366

72 Baig J, Chanut F, Kornberg TB, Klebes A. The chroma-tin-remodeling protein Osa interacts with CyclinE in Drosophila eye imaginal discs. Genetics, 2010, 184: 731–744

73 Zavadil J, Bottinger EP. TGF-beta and epithelial-to-mesenchymal transitions. Oncogene, 2005, 24: 5764–5774

Open Access This article is distributed under the terms of the Creative Commons Attribution License which permits any use, distribution, and reproduction

in any medium, provided the original author(s) and source are credited.

Supporting Information

Table S1 Primary RNAi screen results and signaling classification of the candidate genes identified by the RNAi screen.

The supporting information is available online at life.scichina.com and link.springer.com. The supporting materials are

published as submitted, without typesetting or editing. The responsibility for scientific accuracy and content remains entirely with the authors.